Part2Vec part vectorization processing method based on deep learning

Abstract

The invention discloses a part vectorization processing method based on deep learning, namely Part2Vec, which mainly comprises the following steps of preprocessing an existing product structure, creating a part dictionary, and establishing a part dictionary based on a constraint relationship among parts; constructing a training sample and a sample set, determining a neural network structure, training obtained sample set data, obtaining a weight matrix between an input layer and a hidden layer, converting the weight matrix into an embedded matrix, wherein the creation sequence and the description sequence are the same. According to the method, vectorized representation can be quickly carried out on the parts, so that the unique representation of the parts is realized, the correlation between similar parts is reserved, and the dimensionality of the parts during numeralization is reduced.

Description

technical field [0001] The present invention relates to a vectorization processing method, in particular to a Part2Vec part vectorization processing method based on deep learning. Background technique [0002] With the continuous deepening of intelligent design, higher requirements are put forward for product design. How to quickly carry out intelligent product design, including intelligent reuse and intelligent configuration, has become one of the hot spots and problems that many scholars are keen to study. [0003] In the process of intelligent design, the most critical difficulty lies in how to effectively represent the parts. Some methods are usually used, such as One-Hot vector, PF-IPF (for the specific process, please refer to the patent "calculation of product structure similarity based on TF-IDF idea method"), etc., to vectorize the parts. [0004] However, when using One-Hot vectors to represent parts, the inner products of One-Hot vectors between parts are all zer...

AI Technical Summary

Problems solved by technology

[0004] However, when using One-Hot vectors to represent parts, the inner products of One-Hot vectors between parts are all zero, which is easy to cause the isolated state of parts, ignoring the correlation between parts, and the difficulty of network training will vary with the dimension of One-Hot vectors. PF-IPF is used to represent parts, and the importance of a part is simply measured by "part frequency". The part with a high frequency is less important, while the part with a small frequency is more important. This is obviously not applicable to In some special cases, there are many limitations in the use of this method

Images